Positioning systems based on signal fingerprinting, e.g., RF signal matching, rely on fingerprint prediction, e.g., predicting what RF signals will be detectable at a particular location. The prediction may be represented using a map of the geographic region over which the positioning system operates. The map may contain information describing the spatial (positional) variation of the physical characteristics of the signal(s) that constitute the fingerprint, e.g., RF prediction map, or some portion thereof. In many cases the accuracy of the position estimate produced by a positioning algorithm using a fingerprint map depends upon the accuracy of the underlying fingerprinting map.
Fingerprint maps, e.g., RF prediction maps, may be obtained by various methods. One approach is to determine the fingerprint maps using parametric models of the propagation of the physical signal used by the positioning system. For example, in the case of RF signal based positioning, a ray-tracing algorithm may be used to compute an estimate of received signal power or signal propagation time for a given transmit location. In the indoor setting, e.g., inside a building, the propagation predicted by the ray-tracing algorithm may depend on a number of parameters, such as the RF properties of the wall/ceiling/floor materials, including reflection, refraction and diffraction effects. One benefit of using a parametric model is that the parameters, together with the building map, constitute a compact representation of the information used to estimate signal characteristics for arbitrary transmit and receive locations. If the number of potential transmit and receive locations is large, then such a model can be a more efficient means of data representation as compared to, say, directly storing the fingerprint value as a function of transmit and receive position. Efficient representation implies smaller storage requirements and/or more efficient communication of map information. Communication of map information often plays an important role in positioning systems.
One problem with the parametric model based approach to generating fingerprint maps is that the reliability of the map depends on the accuracy of the parameters used in generating the map. While the parameters may be based on an initial RF signal survey made at a point in time or over some time period, after the survey is completed RF signal sources may change and/or physical conditions may change affecting RF signal propagation at one or more locations covered by a fingerprint map. It can be costly and/or impractical to conduct RF surveys frequently.
In view of the above discussion, it should be appreciated that there is a need for methods and apparatus which can be used in updating parameters used to generate prediction maps, e.g., without requiring a complete RF survey to be conducted as part of the updating process.